Reconfigurable automatic driving trajectory tracking fault-tolerant motion control method and system for multi-axle commercial vehicle

Abstract

The invention belongs to the technical field of commercial vehicle automatic driving, and discloses a multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control method, which specifically comprises the following steps: establishing a vehicle reconfigurable dynamic tracking error prediction model based on a small corner hypothesis and a vectorization modeling method; augmenting a state quantity for a delay phenomenon of a vehicle steering system; traversing different vehicle steering failure configurations offline to generate a steady state reference and a control reference; system feedback is obtained online in real time; an MPC problem with equality constraints is converted into a quadratic programming problem for solving and calculation simplification; performing matrix multiplication on the optimal control law obtained by looking up the table according to the time-varying parameters and state feedback to obtain a steering wheel angle control quantity; and sending the turning angle instruction to a steering wheel of the multi-axle vehicle. According to the invention, the compatibility and real-time performance of the automatic driving motion controller of the multi-axle commercial vehicle can be improved, and the precision of the trajectory tracking process can be improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of automatic driving of commercial vehicles, and particularly relates to a multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control method and system. BACKGROUND

[0002] The design of a multi-axle vehicle can support the installation of multiple steering actuators, thereby achieving more flexible and stable motion control. However, the increase in the number of steering actuators also brings a higher failure rate, which puts higher requirements on the safety of an autonomous vehicle. At the same time, with the increase in the number of actuators, the complexity of the system in the state and control dimensions is also increased, which brings certain challenges to the computing power and real-time response capability of the controller. This paper aims to solve the above problems and invent a unified reconfigurable fault-tolerant control strategy applicable to different numbers of axles and steering actuator configurations.

[0003] Although existing research can accurately model vehicles of different structures by coordinating the position of the tires, the modeling process is simplified to some extent. However, these methods have not fully utilized the advantages of reconfigurable dynamic models, and further consideration of the response delay of steering actuators and their fault tolerance is needed.

[0004] Steering delay has a significant impact on high-speed motion control of autonomous vehicles. Existing research shows that when the steering system has a delay of more than 50 ms, a steer-by-wire system using proportional-derivative control will lose stability. Existing research handles the delay of the steering system through a state observer or a feedforward controller, but this often does not utilize the control horizon in model predictive control methods. The added state observer and feedforward controller will occupy a large amount of computing resources of the chassis domain controller.

[0005] Multi-axle commercial transport vehicles, closed port automatic driving board vehicles, and mine dump trucks, etc. heavy engineering and logistics transport vehicles, are prone to steering actuator failure, degradation, and failure during long-term operation. When a steering actuator fails, it cannot provide sufficient lateral force for the vehicle, thereby losing the ability to maintain trajectory tracking during steering. It is extremely important to maintain the steering ability of the vehicle using the remaining steering actuators before the automatic driving system is downgraded.

[0006] Multi-axle vehicles allow more steering actuators and drive actuators to be installed, thereby achieving more flexible and stable motion control. However, with the increase in the number of actuators, the state dimension and control dimension of the system are also increased, and it is challenging to solve the control sequence of this large-dimensional system in real time on the chassis domain controller according to the control period. SUMMARY

[0007] In view of the problems existing in the prior art, the application provides a multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control method.

[0008] The application is implemented as follows: a multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control method comprises the following steps:

[0009] Step 1: based on the small angle assumption and the vectorization modeling method, the position and heading relationship between the vehicle and the planned trajectory point is considered, and a vehicle reconfigurable dynamics tracking error prediction model is established;

[0010] Step 2: the delay phenomenon of the vehicle steering system is modeled in the form of a pure time delay and a first-order inertia element in series, and the state quantity of the first-order inertia element is augmented into the state equation;

[0011] Step 3: under different steering actuator error configuration conditions, the minimum tire side slip angle of the vehicle is taken as the optimization objective, the vehicle prediction model and the position error and heading error of the trajectory point are taken as the equality constraints, an optimization problem is established, and a steady-state reference is solved;

[0012] Step 4: the lateral velocity and the yaw rate of the vehicle chassis feedback are obtained in real time, and the lateral error and the heading angle error of the vehicle and the target trajectory point are calculated;

[0013] Step 5: the MPC problem with equality constraints is converted into a quadratic programming problem for solving and calculation simplification;

[0014] Step 6: for the vehicle prediction model, the steady-state reference optimization solving and the quadratic programming problem with equality constraints established above, time-varying parameters are traversed for offline solving, and are stored according to a certain rule;

[0015] Step 7: the optimal control law obtained according to the time-varying parameter lookup table and the state feedback are subjected to matrix multiplication to obtain the steering wheel steering angle control quantity;

[0016] Step 8: the steering wheel of the multi-axle vehicle is controlled to execute the steering angle instruction.

[0017] Further, the step 1 and the step 2 comprise:

[0018] The state equation of the reconfigurable vehicle prediction model after the state quantity of the first-order inertia delay of the steering system is augmented is as follows:

[0019]

[0020] Further, the step 3 comprises:

[0021] Under different steering actuator failure configurations, the equality constraints of the state equation and the elimination of the heading angle error will be different, and a unified reference generator is established by considering different failure conditions as follows:

[0022]

[0023] Further, the step 6 and step 7 include:

[0024] The fault-tolerant MPC problem is converted into the following quadratic programming (QP) problem for solving, in the form as follows:

[0025]

[0026] s.t.A I z=b

[0027]

[0028] The KKT equation of the QP problem is:

[0029]

[0030] The optimal control quantity obtained by solving can be expressed as:

[0031]

[0032] The symbols and numbers in the above formula are explained: v x is the longitudinal speed of the vehicle; r is the yaw rate of the vehicle; m is the mass of the vehicle; I z is the moment of inertia of the vehicle; is the lateral force vector of the tire;, is the coordinate of the wheel center in the vehicle coordinate system; is the tire cornering stiffness vector; p δ is the steering reconfigurable matrix; d s is the actual steering angle of the steering wheel; d c is the steering delay front angle of the vehicle; K is the road curvature of the planning trajectory point.

[0033] Another purpose of the present application is to provide a multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control system, comprising:

[0034] The reconfigurable dynamic model is established, which is used to establish a vehicle reconfigurable dynamic tracking error prediction model based on the small angle assumption and the vectorization modeling method, considering the position and heading relationship between the vehicle and the planning trajectory point;

[0035] The modeling of the vehicle steering system delay is used to model the delay phenomenon of the vehicle steering system in the form of a pure time delay and a first-order inertia element in series, and the state quantity of the first-order inertia element is augmented into the state equation;

[0036] A high-stability margin control quantity solution considering fault tolerance is used to consider different steering actuator error configurations, minimize the tire side slip angle of the vehicle as the optimization target, and establish an optimization problem with the vehicle prediction model and the position error and heading error of the trajectory point as the equality constraint to solve the steady-state reference;

[0037] Chassis state feedback is used to obtain the lateral velocity and yaw angular velocity of the vehicle chassis feedback online in real time, and calculate the lateral error and heading angle error of the vehicle and the target trajectory point;

[0038] Offline solution simplifies the calculation, which is used to convert the MPC problem with equality constraints into a quadratic programming problem for solution and calculation simplification;

[0039] Offline KKT matrix MAP storage is used to traverse the time-varying parameters of the vehicle prediction model, steady-state reference optimization solution and quadratic programming problem with equality constraints for offline solution, and stored according to a certain rule;

[0040] Online steering control quantity calculation is used to obtain the steering wheel angle control quantity by matrix multiplication of the optimal control law obtained by looking up the table according to the time-varying parameters and the state feedback;

[0041] Another object of the present application is to provide a computer device comprising a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to make the processor execute the steps of the multi-axle commercial vehicle reconstructable automatic driving trajectory tracking fault-tolerant motion control method.

[0042] Another object of the present application is to provide a computer readable storage medium storing a computer program, the computer program being executed by a processor to make the processor execute the steps of the multi-axle commercial vehicle reconstructable automatic driving trajectory tracking fault-tolerant motion control method.

[0043] Another object of the present application is to provide an information data processing terminal for implementing the multi-axle commercial vehicle reconstructable automatic driving trajectory tracking fault-tolerant motion control system.

[0044] In combination with the above technical solutions and the technical problems solved, the technical solutions protected by the present application have the following advantages and positive effects:

[0045] Firstly, the reconstructable vehicle dynamics model proposed in the present application fully considers the characteristics of different configurations of commercial vehicles with variable number of axles, and adopts a vectorized modeling method, which can adapt to multi-axle vehicles with different numbers of steering actuators.

[0046] The application utilizes the advantages of the prediction time domain and control time domain of the model predictive control method to process the delay existing in the steering system, and considers all error configurations of the vehicle steering actuator in combination with the advantages of offline optimization calculation, so that the fault tolerance is considered when the control quantity is solved online.

[0047] The offline optimization and online lookup table method provided by the application completes the matrix splicing and decomposition work offline, which occupies a large amount of computing resources, improves the execution efficiency of the automatic driving motion controller, and the saved computing resources can be used for the calculation of other actuator control of the chassis domain controller.

[0048] Second, the expected income and business value of the technical scheme of the application after transformation are:

[0049] The reconfigurable automatic driving trajectory tracking fault-tolerant motion control provided by the application can adapt to different vehicle configurations of different axis numbers, that is, it can adapt to new vehicle models developed by vehicle manufacturers, and can reduce the cost consumption of the manufacturers for controller development and verification for new vehicle models.

[0050] The technical scheme of the application fills the technical gap in the industry at home and abroad:

[0051] 1. Reconfigurable modeling of commercial vehicle dynamics model. The vectorized reconfigurable vehicle dynamics modeling constructed by the application can adapt to different commercial vehicle configurations under the condition of changing static parameters, so in the development process of the motion controller of the new vehicle model, it is not necessary to re-perform dynamics modeling for a vehicle with a special configuration, which can reduce the development cycle of vehicle dynamics modeling and shorten the development and research cycle of the automatic driving motion controller of the new vehicle model.

[0052] 2. Fault-tolerant control of commercial vehicle automatic driving lateral motion. The automatic driving lateral motion controller of the application is also reconfigurable, and considers the fault-tolerant control under different error configurations of the vehicle steering actuator, which provides protection for the safety of the vehicle before the system degradation protection when the steering actuator fails in the long-distance high-speed automatic driving working condition of the multi-axis commercial vehicle.

[0053] 3. High real-time reconfigurable automatic driving motion controller. The automatic driving trajectory tracking fault-tolerant controller provided by the application performs offline matrix splicing and inverse operation which occupies a large amount of computing resources, and only needs to calculate the multiplication of the matrix online, which realizes the real-time performance of 0.2ms on the mass-produced vehicle controller. BRIEF DESCRIPTION OF DRAWINGS

[0054] Figure 1 is a reconfigurable automatic driving trajectory tracking fault-tolerant motion control method flowchart of a multi-axis commercial vehicle provided by the embodiment of the application;

[0055] Figure 2A structure block diagram of a multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control system is provided by an embodiment of the present application.

[0056] Figure 3 A method and a test framework diagram are provided by an embodiment of the present application.

[0057] Figure 4 A vehicle reconfigurable dynamics model diagram is provided by an embodiment of the present application.

[0058] Figure 5 A road deviation diagram is provided by an embodiment of the present application.

[0059] Figure 6 A typical multi-axle commercial vehicle structure configuration and an independent steering actuator arrangement mode of front and rear axles thereof are provided by an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present application clearer and more apparent, the present application will be further described in detail below in combination with embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0061] In the automatic driving process of a multi-axle commercial vehicle, due to the large number of axles, significant lateral force coupling of tires, and possible asymmetric damage or performance degradation of steering actuators, the accuracy of the conventional dynamics model based on fixed structure assumptions decreases when performing path tracking tasks. Most existing models ignore the dynamic lag of the steering system and the influence of actuators on the overall path control performance under different fault configurations, resulting in the inability to dynamically reconfigure in time when deviations or actuator failures occur, a decrease in path tracking accuracy, and even the initiation of trajectory divergence, affecting the safety and reliability of automatic driving.

[0062] The modeling framework proposed in the present application is based on the small angle assumption, combined with the vectorization model construction method, and realizes the reconfigurable design of the state prediction model by introducing the relative relationship between the current attitude of the vehicle and the planned trajectory point (including the lateral error and the heading angle error). This modeling method not only retains the computational efficiency advantage of the traditional linear model, but also introduces the first-order inertia response and pure delay link of the steering actuator through the state variable augmentation mechanism, thereby systematically considering the contribution of the actuator dynamic process to the response delay of trajectory tracking.

[0063] In the control design, the core of the method is the dynamic calculation of the steady-state target. Unlike the conventional MPC design which directly adopts the feedforward control target or the fixed weight adjustment strategy, the method reconstructs the tire side slip angle minimization problem in each control period, converts the path tracking performance into a controlled variable optimization problem, and reduces the stability degradation caused by the accumulation of steering angle error. The constraint conditions not only involve the consistency of the state transition model, but also include the suppression target of the heading error and the lateral error estimated in real time, so as to guarantee the geometric convergence characteristics of the control system.

[0064] In order to adapt to the dual requirements of real-time performance and fault tolerance in industrial deployment, the control problem is further converted into a quadratic programming problem with equality constraints, and an offline time-varying parameter traversal mechanism is used to construct a multi-dimensional lookup table structure. Considering the predicted response of the vehicle under different speed, path curvature and actuator fault configuration combinations, the linear mapping coefficient matrix of the optimal control law is constructed in advance through the KKT solving framework, so as to simplify the complex QP solving process into a lookup table and matrix multiplication, and realize the online response of millisecond-level control frequency.

[0065] In actual execution process, the system collects the yaw rate and lateral velocity feedback of chassis sensors, and estimates the current error state in real time through the trajectory tracking error calculation module. Then, the optimal control law mapping table is called in combination with the current time-varying working condition parameters to complete the integrated state feedback control law generation. The system outputs the steering angle command for controlling each available steering actuator. In the presence of some steering shaft failures, the control system realizes the degradation maintenance of the trajectory tracking function through the redistribution of redundant degrees of freedom, and has fault tolerance robustness.

[0066] According to the above analysis, it can be seen that the method not only makes up for the shortcomings of the prior art in steering actuator modeling and multi-configuration redundancy control, but also greatly improves the industrial landing feasibility through the engineering-friendly MPC reconstruction strategy and lookup table mechanism, and provides a scalable high-robustness motion control solution for the automatic driving and trajectory precision control of multi-axle heavy vehicles.

[0067] As shown in Figure 1 The multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control method provided by the embodiment of the application comprises the following steps:

[0068] S101, based on the small angle assumption and the vectorization modeling method, considering the position and heading relationship between the vehicle and the planned trajectory point, a vehicle reconfigurable dynamics tracking error prediction model is established;

[0069] S102, the delay phenomenon of the vehicle steering system is modeled in the form of pure delay and first-order inertia element in series, and the state quantity of the first-order inertia element is augmented into the state equation;

[0070] S103, considering different steering actuator failure configuration cases, taking the minimum tire side slip angle of the vehicle as the optimization target, taking the vehicle prediction model and the position error and heading error of the trajectory point as the equation constraint, establishing an optimization problem for steady-state reference solving;

[0071] S104, obtaining the lateral velocity and yaw rate of the vehicle chassis in real time online, calculating the lateral error and heading angle error of the vehicle and the target trajectory point;

[0072] S105, converting the MPC problem with equation constraints into a quadratic programming problem for solving and calculation simplification;

[0073] S106, for the vehicle prediction model, steady-state reference optimization solving and quadratic programming problem with equation constraints established above, traversing the respective time-varying parameters to perform offline solving, and storing according to a certain rule;

[0074] S107, according to the optimal control law and state feedback obtained by looking up the time-varying parameters, performing matrix multiplication to obtain the steering wheel steering angle control amount;

[0075] S108, controlling the steering wheel of the multi-axle vehicle to execute the steering angle instruction.

[0076] The S101 and S102 provided by the embodiment of the application include:

[0077] The state equation of the reconfigurable vehicle prediction model after the augmented steering system first-order inertia delay state quantity is established as:

[0078]

[0079]

[0080] The S103 provided by the embodiment of the application includes:

[0081] Under different steering actuator failure configurations, the equation constraint of the state equation and the elimination of the heading angle error will be different, considering different failure conditions, a unified reference generator is established as:

[0082]

[0083] The S106 and S107 provided by the embodiment of the application include:

[0084] The fault-tolerant MPC problem is converted into the following quadratic programming (QP) problem for solving, and the form is as follows:

[0085]

[0086] s.t.A I z=b

[0087]

[0088] The KKT equation of the QP problem is:

[0089]

[0090] The optimal control quantity can be obtained by solving:

[0091]

[0092] The symbols and numbers in the above formula are explained as follows: x is the longitudinal speed of the vehicle; r is the yaw rate of the vehicle; m is the mass of the vehicle; I z is the moment of inertia of the vehicle; is the lateral force vector of the tire;, is the coordinate of the wheel center in the vehicle coordinate system; is the tire cornering stiffness vector; p δ is the steering reconfigurable matrix; d s is the actual steering angle of the steering wheel; d c is the steering delay angle of the vehicle; k is the road curvature of the planned trajectory point.

[0093] As Figure 2 shown, the multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control system provided by the embodiment of the application comprises:

[0094] A prediction model establishing module is configured to construct a vehicle state prediction model comprising a lateral speed, a yaw rate, a lateral error and a heading angle error based on a current speed, a posture and a relative relationship of a trajectory point of the vehicle;

[0095] A delay augmentation module is configured to model a pure delay and a first-order inertia element in series for a response process of a steering system of the vehicle, and to augment an inertia state variable to a state vector of the prediction model;

[0096] A reference solving module is configured to solve a steady-state control target with a minimum tire cornering angle based on current state estimation and trajectory point information under different actuator configurations and fault scenarios, and to output an expected control input as a model predictive control reference;

[0097] A feedback calculation module is configured to collect a yaw rate and a lateral speed of a chassis feedback of the vehicle in real time, and to calculate a lateral error and a heading angle error of the vehicle in combination with a planned path;

[0098] A quadratic programming solving module is configured to convert a model predictive control problem into a quadratic programming problem with equality constraints under time-varying states, and to call a control law generation module to return a control quantity solution;

[0099] A parameter storage module is configured to call an optimal control law matrix or KKT mapping result corresponding to a current time-varying parameter in a lookup table structure based on a current working condition parameter of the vehicle, including a longitudinal speed, a path curvature, and a steering delay time; a control law generation module is configured to receive a reference control amount and a current state amount, and generate a final control input vector based on a lookup table result of the parameter storage module through matrix multiplication, and output the final control input vector to an execution module;

[0100] An execution module is configured to convert a control amount output by the control law generation module into a specific multi-axle vehicle front and rear wheel steering angle control instruction, and drive each axle steering actuator to perform an action.

[0101] In the multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control system, signal data processing first obtains chassis dynamic information and environment perception information from a vehicle sensor network, including a longitudinal speed, a lateral speed, a yaw rate, a vehicle attitude angle, and a path point sequence output by a trajectory planning module. After time synchronization and coordinate system conversion, the sensor data is input to a state estimation unit, combined with Kalman filtering or extended Kalman filtering algorithm, to filter out measurement noise and interference, and output a high-precision vehicle dynamic state vector as a basic input for prediction model establishment and feedback calculation.

[0102] In the delay augmentation module, the system models the dynamic characteristics of the steering actuator, equivalent to a state space expression by connecting a pure delay element and a first-order inertia element in series, and adds the inertia state variable inside the actuator to the extended state vector. This augmentation process enables the prediction model to explicitly consider the dynamic lag effect of the steering system during calculation, avoiding distortion of the control law for future state estimation, and providing a more realistic system model for subsequent quadratic programming solution.

[0103] The reference solving module calculates a steady-state control target in real time according to the current state estimation result and the target trajectory point information. In the signal processing process, the module converts the trajectory point curvature information into expected vehicle yaw rate and lateral acceleration reference values, and solves the minimum side slip angle distribution scheme through a tire side slip stiffness model. This solving process is self-adaptive under different actuator configurations and fault scenarios to ensure that the control reference signal meets the trajectory tracking requirements and suppresses the influence of excessive tire slip on vehicle stability.

[0104] The quadratic programming solution module receives the delayed augmented prediction model and the reference control target, unifies the trajectory tracking error, the control input rate of change and the actuator constraint condition into a time-varying constraint optimization problem, and converts it into a quadratic programming form with an equality constraint by using the KKT condition. At the signal processing level, the core calculation process of the module includes state prediction vector construction, cost function gradient and Hessian matrix generation, and real-time update of the constraint matrix, and finally outputs the optimal control input solution vector that meets the constraint condition.

[0105] The control law generation module cooperates with the parameter storage module to obtain the optimal control law matrix or the KKT mapping result corresponding to the current longitudinal velocity, path curvature and steering delay through table lookup. The signal data mainly undergoes matrix multiplication operation in this stage, and the reference input and the current state deviation are mapped into the steering angle instruction vector of each axis, and are sent to the execution module. The execution module amplifies, filters and converts the received digital control signal into a driving signal, and finally drives the front and rear wheel steering actuators of the multi-axle vehicle to realize the trajectory tracking closed-loop control under fault-tolerant conditions.

[0106] In the task of reconfigurable commercial vehicle automatic driving trajectory tracking control, as shown in the framework diagram of the application, Figure 3 According to the trajectory point information given by the planning layer, the distance error and the heading error between the trajectory point and the target trajectory point are calculated, the motion state quantity of the chassis and the error configuration of the steering actuator are combined to augment the state feedback, the KKT inverse matrix and the steady-state reference are extracted from the offline optimization solution table, and the corresponding matrix multiplication is performed to obtain the steering angle of each wheel, and the steering angle instruction is sent to the steering actuator.

[0107] Figure 4 Vehicle reconfigurable dynamics model diagram

[0108] The multi-axle commercial vehicle reconfigurable automatic driving trajectory tracking fault-tolerant motion control method provided by the application specifically comprises the following steps:

[0109] Step 1: The vehicle prediction model is reconfigured by using the vectorization method. Appropriate assumption conditions are proposed, and the balance equations in the lateral and yaw directions of the vehicle are listed according to Newton-Euler law to obtain:

[0110]

[0111] Where, v x and v y are the longitudinal and lateral speeds of the vehicle, r is the yaw angular velocity of the vehicle, m and I z are the mass and moment of inertia of the vehicle, (n t is the number of wheels) is the lateral force of the tire, is the coordinate of the wheel center in the vehicle coordinate system, and denotes the element-wise multiplication and addition operation.

[0112]

[0113] δ v = p δ δ s

[0114]

[0115] where, is the steering reconfigurable matrix; n δ is the number of steering actuators, assuming p δ [i,j] is zero; represents the steering angle of the ith tire and the steering angle of the jth actuator, and otherwise p δ [i,j] represents the steering ratio; is the actual steering angle of the tire; is the actual steering angle of the steering actuator; The time delay of the steering angle of the vehicle is considered; is the time constant of the steering system, and τ d is the pure time delay.

[0116] Step two, establish the tracking error model and arrange the state equation. As shown in Figure 5 , the lateral error is defined according to the distance between the vehicle mass center and the desired trajectory point, which is denoted by e y , and the vector with the desired trajectory point as the starting point and the vehicle mass center as the ending point, and the direction pointing to the left side of the road reference point is positive and the right side is negative. e φ represents the heading angle error of the vehicle, and the direction pointing to the driving direction of the vehicle with the desired trajectory point as the starting point is positive counterclockwise and positive clockwise. The corresponding state equation is as follows:

[0117]

[0118] where, κ is the reference curvature.

[0119] The summary in the form of the state equation is as follows:

[0120]

[0121] The state equation is converted into a discrete form by the bilinear transformation method or the zero-order holder (ZOH) method.

[0122] Step three, the solution of the state and control steady-state reference of the system. The state and control reference is very important. However, the reference generator of the controller design is unified under different actuator fault conditions:

[0123]

[0124] Generally, the vehicle can keep the heading angle error as 0 in steady state when full wheel steering 1 in full wheel steering case, and 0 otherwise. In the actuator fault configuration and fixed vehicle speed case, the above optimization problem is a typical parameter-varying quadratic programming problem, whose parameter is v x and the actuator fault configuration. Meanwhile, the actuator fault configuration number is obtained by converting the binary array composed of the diagonal elements of the actuator fault Boolean matrix p into a decimal number.

[0125] Step 4, MPC problem establishment and solution. The multi-axle reconfigurable vehicle model predictive trajectory tracking control problem is converted into the following quadratic programming (QP) problem for solution, which is as follows:

[0126]

[0127] s.t.A I z=b

[0128] wherein,

[0129]

[0130] Let the decision variable be wherein z i =[x i ,u i ].

[0131]

[0132] wherein, in the above formula,

[0133]

[0134] The equality constraints composed of the state equation and the control delay are as follows:

[0135]

[0136] wherein, is the state feedback vector, is the control delay vector, which is as follows:

[0137]

[0138] The KKT equation of the above QP problem can be expressed as:

[0139]

[0140] Written as a matrix:

[0141]

[0142] Solving the above equation can obtain:

[0143]

[0144] For time-invariant system, although there are disturbance changes, because the disturbance does not affect K * , and when the weight parameter of the MPC problem is fixed, K * is a constant, so it can be calculated offline.

[0145] Step five, reduce the online calculation amount of MPC. According to the MPC solving process established in the above steps, the following matrix relationship can be constructed:

[0146]

[0147] Wherein,

[0148] The calculation result of the control amount can be changed as follows:

[0149]

[0150] Wherein, Although the controller considers the delay and fault reconstruction control problem, the optimal control rate is simplified to matrix vector multiplication operation. Because the Therefore, the adjacent point interpolation needs to be used in the configuration, and the linear interpolation needs to be used in the speed. That is, the calculation complexity is two matrix vector multiplication operations, and the calculation complexity changes linearly with N and n u .

[0151] Step six, offline optimization solving. The steady-state reference calculation established in the above steps is traversed according to the set lateral vehicle speed, longitudinal vehicle speed, yaw rate, and stored according to a certain rule, and the corresponding state reference and control reference are obtained by online multidimensional interpolation. According to the vehicle speed range and all error configurations, the MPC is solved offline, and the optimal control law is stored according to a certain rule.

[0152] Step seven, obtain the state feedback of the system. Accept the eastward speed, northward speed, yaw rate, heading angle, longitude, and latitude information sent by the combined inertial navigation through the CAN bus, and perform coordinate conversion to obtain the vehicle lateral coordinate, longitudinal coordinate, lateral speed, longitudinal speed, yaw rate, and heading angle information in the vehicle coordinate system.

[0153] Step eight, obtain the target trajectory point deviation information. The real-time horizontal and vertical coordinates and the heading angle in the vehicle coordinate system and the coordinate information of the target trajectory point are calculated according to the formula in step two to obtain the lateral deviation and the heading deviation, which are fed back to the MPC controller.

[0154] Step nine, the chassis feedback quantity and the tracking error feedback quantity in the above steps are multiplied by the optimal control law, the steady state reference and the control reference obtained according to the current time-varying reference according to the calculation method in step five to obtain the steering control quantity, which is sent to the steering actuator in real time.

[0155] Figure 6 The structure of the typical multi-axle commercial vehicle and the arrangement of the independent steering actuators of the front and rear axles thereof are provided by the embodiment of the present application.

[0156] It should be noted that the embodiments of the present application can be realized by hardware, software or a combination of software and hardware. The hardware part can be realized by special logic; the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or a specially designed hardware. Those skilled in the art can understand that the above-mentioned devices and methods can be realized by computer executable instructions and / or included in processor control codes, such as carrier media, such as magnetic disk, CD or DVD-ROM, programmable memory, such as read-only memory (firmware), or data carrier, such as optical or electronic signal carrier. The devices of the present application and their modules can be realized by hardware circuit, such as ultra-large scale integrated circuit or gate array, semiconductor, such as logic chip, transistor, etc., or programmable hardware device, such as field programmable gate array, programmable logic device, etc., or by software executed by various types of processors, or by a combination of the above-mentioned hardware circuit and software, such as firmware.

[0157] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any modification, equivalent replacement and improvement within the technical range disclosed by the present application and within the spirit and principle of the present application should be covered within the protection scope of the present application.

Claims

1. A reconfigurable autonomous driving trajectory tracking fault-tolerant motion control method for multi-axle commercial vehicles, characterized in that, include: Step 1: Construct a vehicle dynamics tracking error prediction model under the small turning angle assumption; Step 2 involves modeling the pure time delay of the steering system in series with the first-order inertial element and augmenting the first-order inertial state to the state equation. Step 3 aims to minimize the tire slip angle and solves for the steady-state reference under equality constraints; Step 4: Obtain lateral velocity and yaw rate in real time and calculate lateral error and heading error; Step 5 transforms the model predictive control problem into a quadratic programming problem; Step 6: Solve for and store the optimal control law offline by iterating through the time-varying parameters. Step 7: Calculate the steering wheel angle control value based on the parameter lookup table and status feedback; Step 8 outputs the steering angle command and drives the steering wheel.

2. The method according to claim 1, characterized in that, The state vector of the prediction model established in step 1 includes the vehicle's lateral velocity, vehicle's yaw rate, vehicle's lateral position error, vehicle's heading angle error, and the inertial state of the steering system.

3. The method according to claim 1, characterized in that, Step 3 obtains the steady-state reference control quantity by eliminating lateral and heading errors and constraining the tire slip angle.

4. A reconfigurable autonomous driving trajectory tracking fault-tolerant motion control system for multi-axle commercial vehicles, characterized in that, include: The prediction model building module is used to construct a vehicle state prediction model that includes lateral velocity, yaw rate, lateral error and heading angle error based on the relative relationship between the vehicle's current speed, attitude and trajectory points. The delay augmentation module is used to perform pure time delay and first-order inertial element concatenation modeling of the response process of the vehicle steering system, and to augment the inertial state variables into the state vector of the prediction model. The reference solver module is used to solve the steady-state control objective with the minimum tire slip angle based on the current state estimate and trajectory point information under different actuator configurations and fault scenarios, and output the desired control input as a model predictive control reference. The feedback calculation module is used to collect the yaw rate and lateral rate fed back by the vehicle chassis in real time, and calculate the vehicle's lateral error and heading angle error in combination with the planned path. The quadratic programming solution module is used to transform the model predictive control problem into a quadratic programming problem with equality constraints under time-varying conditions, and call the control law generation module to return the control quantity solution. The parameter storage module is used to retrieve the optimal control law matrix or KKT mapping result corresponding to the current time-varying parameters from the lookup table structure based on the vehicle's current operating parameters, including longitudinal speed, path curvature, and steering delay time. The control law generation module is used to receive the reference control quantity and the current state quantity, and generate the final control input vector through matrix multiplication based on the lookup result of the parameter storage module, and output it to the execution module. The execution module is used to convert the control quantity output by the control law generation module into specific steering angle control commands for the front and rear wheels of a multi-axle vehicle, and drive the steering actuators of each axle to perform actions.

5. The system according to claim 4, characterized in that, The delay augmentation module employs a pure time delay and a first-order inertial element in series structure, and writes the inertial state into the state vector.

6. The system according to claim 4, characterized in that, The parameter storage module stores the optimal control law using vehicle longitudinal speed, road curvature, and steering delay as indexes.

7. A multi-axle commercial vehicle, characterized in that, The vehicle includes the control system of claim 4 and a plurality of electric steering actuators, and is able to maintain trajectory tracking stability in the event of actuator failure.

8. The vehicle according to claim 7, characterized in that, When a steering actuator malfunction is detected, the control system shuts down the malfunctioning actuator and redistributes the steering angle of the remaining actuators to reduce the tire slip angle.

9. A computer-readable storage medium having instructions stored thereon, which, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 3.

10. A computer device comprising a memory and a processor, characterized in that, The memory stores instructions that, when executed by a processor, cause the computer device to perform the method according to any one of claims 1 to 3.

Images